Utilizing Predictive Modeling in Songbird
Predictive modeling in Songbird uses advanced machine learning (ML) models to forecast the performance of your music tracks. These models, integrated through Songbird’s internal tooling, offer valuable insights into streaming success, listener engagement, and financial forecasting.
Here’s a detailed guide on how to use each predictive model effectively.
General Requirements for All Predictive Models
Before diving into each model, it’s important to note the common requirements:
- Artist Chartmetric ID: Essential for all models to associate the track with the correct artist. If you need assistance obtaining this ID or adding your artist data to Songbord, please contact support@unbias.co.
- Audio File: Each model requires the upload of an audio file to analyze the track’s potential.
- Supported Formats: Ensure your audio file is in MP3, M4A, or WAV format.
Pre-Release Prediction
The Pre-Release Prediction model estimates streaming milestones for unreleased songs, helping artists and labels plan their marketing and promotional strategies.
Steps to Use the Pre-Release Prediction Model
-
Click the Prediction button
-
Select Pre-Release Prediction: Choose this option from the available models.
-
Input Artist Chartmetric ID: Enter the ID to link the track with its artist.
-
Upload Your Track: Provide the audio file of your unreleased song.
-
Submit for Analysis: Once all information is inputted, click ‘Analyze’.
-
Review Predictions: The model will offer forecasts on streaming milestones, aiding in setting achievable targets.
Post-Release Forecast
This model provides a 30-day streaming forecast for songs already available on Spotify, utilizing cumulative stream counts from the past seven days. Note that the song must not be older than two years.
Steps to Use the Post-Release Forecast Model
-
Select Post-Release Forecast: Within the ‘Prediction’ section, find this model and select it.
-
Provide Cumulative Streams: Enter streams the song has accumulated for each day, over the past 7 days. The stream counts are easily available in the artist’s Spotify For Artist dashboard.
-
Input Artist Chartmetric ID and Upload the Track: As with the Pre-Release model, include the artist’s Chartmetric ID and the track’s audio file.
-
Analyze and Review: Submit your data for analysis. The model will return a forecast for the next 30 days.
ROI in Streams Prediction
This model predicts the timeframe within which a track will achieve specific streaming counts, directly correlating to the anticipated return on investment (ROI). For example, if the costs for marketing the track was 1,000 / $0.003)
Steps to Use the ROI in Streams Prediction Model
- Choose ROI in Streams Prediction: In the predictive modeling section, select this option to proceed.
- Input Required Information: Include the artist’s Chartmetric ID and the relevant audio file.
- Specify Desired Stream Count: Input the streaming milestone you’re aiming for with this track.
- Submit for Analysis: With all details provided, click ‘Analyze’ to receive your prediction.
- Interpret Results: The model will offer a timeline prediction for reaching your specified stream count, assisting in financial and promotional planning. `
Following these steps for each predictive model will enable you to leverage Songbird’s machine learning capabilities for detailed performance forecasting. Remember, the accuracy of predictions can be influenced by external factors and market dynamics, so consider these forecasts as part of a comprehensive strategy rather than absolute certainties.